-
Notifications
You must be signed in to change notification settings - Fork 4.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Azure.Core pipeline has a very high failure rate and we need to get this green. #42287
Comments
Per offline discussion, we should not remove tests from the pipeline to achieve this goal, as it allows build breaks to be introduced from core/generator changes that take a lot of time to track down. |
I think one of the things to consider is allowing a pipeline-specific override for global test timeout, so that large suites such as Core and ResourceManager can offset the non-determinism for scheduling delays. |
I looked into this a bit and noticed something interesting. There seems to be a correlation with TestProxy failures in the test logs. In speaking with @scbedd , he thinks this issue could be related. This would address the 500s, but there is still the issue of proxy cloning of recordings causing delays that are not the fault of the test itself. I'm going to investigate if it is feasible to time the proxy initialization phase and subtract that from the test's total time somehow. Alternate approach idea from @scbedd is to batch restore all test assets somehow prior to normal test execution. That seems more involved, but would also address the tests not being charged for proxy setup time. example of proxy failures related to timeouts:
|
Looks like we already subtract the proxy setup time from the test execution time So let's wait to see if fixing Azure/azure-sdk-tools#7784 addresses the majority of cases. |
Hey @christothes I don't know if it totally addressed your issues but your mac builds definitely stabilized a ton after the update! |
The data looks promising! |
Azure.Core pipelines today have a pass rate of about 50% which impedes work and progress as shown below. We need to get this closer to 100%.
https://dev.azure.com/azure-sdk/internal/_pipeline/analytics/stageawareoutcome?definitionId=1180&contextType=build has more information.
The text was updated successfully, but these errors were encountered: